\documentclass[11pt]{article}
\usepackage{float,amsmath,amssymb,amsthm,fullpage,setspace,natbib,enumerate,url,multirow,lscape}
\onehalfspacing
\begin{document}
\begin{table}[htp]\centering\scriptsize
\begin{singlespace}
\begin{tabular}{ l 
c c c c c
 } \hline \hline
\\
 & 2000 & 2004 & 2008 & 2012 & 2016 \\ 

\\
\textbf{General Election:} \\
\hspace{.4cm}Positive GRPs & 570 & 1073 & 1479 & 1479 & 543 \\ 

\hspace{.4cm}Negative GRPs & 848 & 2309 & 2853 & 7816 & 2266 \\ 

\\ \textbf{Primary Election:} \\
\hspace{.4cm}Positive GRPs & 206 & 440 & 2104 & 481 & 2803 \\ 

\hspace{.4cm}Negative GRPs & 26 & 274 & 170 & 733 & 1263 \\ 

\\
\hline \hline
\end{tabular}
\end{singlespace}
\caption{\footnotesize{Positive and Negative Advertising by Year. Numbers reported are measured in 1000s of GRPs.}}
\label{PosNegTable}
\end{table}
\begin{table}[htp]\centering\scriptsize
\begin{singlespace}
\begin{tabular}{ l 
c c c c c c
 } \hline \hline
\\
 & (1) & (2) & (3) & (4) & (5) & (6)
\\ \textbf{DV:} & \multicolumn{3}{c}{Favorability} & \multicolumn{3}{c}{Candidate Choice} \\
 & 2 weeks out & 4 weeks out & 8 weeks out & 2 weeks out & 4 weeks out & 8 weeks out \\ 

 \\
\textbf{Independent Variables:}\\
\hspace{.4cm}Party Match & 0.451*** & 0.460*** & 0.493*** &  &  &  \\ 

 & (0.003) & (0.003) & (0.003) &  &  &  \\ 

\hspace{.4cm}Run Positive Ads & 0.005*** & 0.005** & 0.005** & 0.119*** & 0.069** & 0.094** \\ 

 & (0.001) & (0.002) & (0.002) & (0.016) & (0.024) & (0.032) \\ 

\hspace{.4cm}Run Negative Ads & 0.016*** & 0.017*** & 0.011 & 0.257*** & 0.225*** & 0.095+ \\ 

 & (0.003) & (0.004) & (0.007) & (0.038) & (0.044) & (0.054) \\ 

\hspace{.4cm}Attacked by Negative Ads & 0.001 & 0.004 & 0.010 & 0.091** & 0.109** & 0.080+ \\ 

 & (0.003) & (0.004) & (0.006) & (0.031) & (0.034) & (0.042) \\ 

\\
\textbf{Observations} & 319044 & 306273 & 277205 &  &  &  \\ 

\textbf{Respondents} & 105679 & 101692 & 93084 & 54160 & 42732 & 28740 \\ 

\\
\hline \hline
\end{tabular}
\end{singlespace}
\caption{\footnotesize{The Effects of Ads on Favorability and Candidate Choice. Analyses include respondent and candidate fixed effects (columns 1-3) and candidate fixed effects (columns 4-6). Standard errors are in parentheses and are clustered by respondent in columns 1-3. One star indicates statistical significance at the 5\% level. Two stars indicates statistical significance at the 1\% level. Three stars indicates statistical significance at the 0.1\% level. A plus sign indicates statistical significance at the 10\% level.}}
\label{ResultsTable}
\end{table}
\begin{table}[htp]\centering\scriptsize
\begin{singlespace}
\begin{tabular}{ l 
c c c
 } \hline \hline
\\
 & (1) & (2) & (3)
\\
 & 2 weeks out & 4 weeks out & 8 weeks out \\ \\
\textbf{Relative Vote Share} \\
\textbf{of Attacked Candidate} \\ \\
0 & 0.462$\dagger$ & 0.305$\dagger$ & 0.990 \\ 

 & (0.095) & (0.127) & (0.684) \\ 

0.5 & 0.562$\dagger$ & 0.403$\dagger$ & 1.699 \\ 

 & (0.118) & (0.172) & (1.671) \\ 

1 & 0.715 & 0.592 & 5.986 \\ 

 & (0.184) & (0.296) & (20.382) \\ 

\\
\hline \hline
\end{tabular}
\end{singlespace}
\caption{\footnotesize{Relative Effectiveness of Positive Advertising. Standard errors are in parentheses. $\dagger$ indicates that the coefficient is statistically distinguishable from 1 at the 10\% level.}}
\label{METable}
\end{table}
\begin{landscape}
\begin{table}[htp]\centering\scriptsize
\begin{singlespace}
\begin{tabular}{ l 
c c c c c c c c c
 } \hline \hline
\\
 & (1) & (2) & (3) & (4) & (5) & (6) & (7) & (8) & (9)
\\ \textbf{Sample:} & \multicolumn{3}{c}{All} & \multicolumn{3}{c}{$Pr(Win) <= 0.4$} & \multicolumn{3}{c}{$Pr(Win) <= 0.2$} \\
 & 2 weeks out & 4 weeks out & 8 weeks out & 2 weeks out & 4 weeks out & 8 weeks out & 2 weeks out & 4 weeks out & 8 weeks out \\ 

 \\
\textbf{Independent Variables:}\\
\hspace{.4cm}Probability (baseline) & 0.388*** & 0.368*** & 0.360*** & 0.223*** & 0.214*** & 0.132** & 0.126* & 0.146* & 0.144+ \\ 

 & (0.010) & (0.011) & (0.015) & (0.031) & (0.034) & (0.044) & (0.064) & (0.066) & (0.075) \\ 

\hspace{.4cm}Party Match & 0.016** & 0.016** & 0.014+ & -0.028** & -0.032*** & -0.048*** & -0.002 & -0.004 & -0.094 \\ 

 & (0.005) & (0.006) & (0.008) & (0.009) & (0.010) & (0.013) & (0.028) & (0.028) & (0.058) \\ 

\hspace{.4cm}Run Positive Ads & -0.001 & 0.004 & -0.004 & 0.000 & -0.002 & -0.003 & -0.005 & -0.009 & -0.006 \\ 

 & (0.007) & (0.010) & (0.015) & (0.011) & (0.013) & (0.017) & (0.012) & (0.013) & (0.018) \\ 

\hspace{.4cm}Run Negative Ads & 0.006 & 0.000 & 0.027 & 0.007 & -0.001 & 0.038 & 0.009 & 0.001 & 0.027 \\ 

 & (0.017) & (0.019) & (0.032) & (0.017) & (0.019) & (0.038) & (0.017) & (0.019) & (0.039) \\ 

\hspace{.4cm}Attacked by Negative Ads & 0.031 & 0.062+ & -0.022 & 0.099* & 0.086+ & -0.048 & 0.155** & 0.122* & -0.449 \\ 

 & (0.025) & (0.032) & (0.068) & (0.046) & (0.048) & (0.096) & (0.053) & (0.054) & (0.347) \\ 

\\
\textbf{Observations} & 28805 & 25661 & 18653 & 20326 & 18525 & 14134 & 17361 & 16177 & 12716 \\ 

\textbf{Respondents} & 8843 & 7760 & 5468 & 8149 & 7168 & 5116 & 7220 & 6523 & 4769 \\ 

\\
\hline \hline
\end{tabular}
\end{singlespace}
\caption{\footnotesize{The Effects of Ads on Perceived Probability of Winning the Primary. Analyses include respondent and candidate fixed effects. Standard errors clustered by respondent are in parentheses. One star indicates statistical significance at the 5\% level. Two stars indicates statistical significance at the 1\% level. Three stars indicates statistical significance at the 0.1\% level. A plus sign indicates statistical significance at the 10\% level.}}
\label{ProbTable}
\end{table}
\end{landscape}
\begin{table}[htp]\centering\scriptsize
\begin{singlespace}
\begin{tabular}{ l 
c c
 } \hline \hline
\\
 & \textbf{Gore} & \textbf{Bradley} \\ 

\\
\textbf{Baseline} & 66.90\% & 33.10\% \\ 

\\
\textbf{Positive Ads} (+1000 GRPs) & 1.50\% & 1.54\% \\ 

\\
\textbf{Negative Ads} (+1000 GRPs)\\
\hspace{.3cm}Attack Gore &  & 2.62\% \\ 

\hspace{.3cm}Attack Bradley & 2.51\% &  \\ 

\\
\hline \hline
\end{tabular}
\end{singlespace}
\caption{\footnotesize{Effect Sizes for Advertising (2000 Democratic Primary) -- Analyses report the change in predicted vote share from increasing positive or negative ad spending by 1000 GRPs, relative to a baseline of no ad spending by all candidates.}}
\label{EffectTable1}
\end{table}
\begin{table}[htp]\centering\scriptsize
\begin{singlespace}
\begin{tabular}{ l 
c c c c c
 } \hline \hline
\\
 & \textbf{Bush} & \textbf{McCain} & \textbf{Forbes} & \textbf{Keyes} & \textbf{Bauer} \\ 

\\
\textbf{Baseline} & 63.78\% & 27.12\% & 4.26\% & 3.92\% & 0.93\% \\ 

\\
\textbf{Positive Ads} (+1000 GRPs) & 1.57\% & 1.38\% & 0.29\% & 0.27\% & 0.07\% \\ 

\\
\textbf{Negative Ads} (+1000 GRPs)\\
\hspace{.3cm}Attack Bush &  & 2.62\% & 0.66\% & 0.61\% & 0.15\% \\ 

\hspace{.3cm}Attack McCain & 3.22\% &  & 0.86\% & 0.79\% & 0.20\% \\ 

\hspace{.3cm}Attack Forbes & 4.73\% & 4.52\% &  & 0.92\% & 0.23\% \\ 

\hspace{.3cm}Attack Keyes & 4.75\% & 4.53\% & 0.99\% &  & 0.23\% \\ 

\hspace{.3cm}Attack Bauer & 4.95\% & 4.63\% & 1.01\% & 0.93\% &  \\ 

\\
\hline \hline
\end{tabular}
\end{singlespace}
\caption{\footnotesize{Effect Sizes for Advertising (2000 Republican Primary) -- Analyses report the change in predicted vote share from increasing positive or negative ad spending by 1000 GRPs, relative to a baseline of no ad spending by all candidates.}}
\label{EffectTable2}
\end{table}
\begin{landscape}
\begin{table}[htp]\centering\scriptsize
\begin{singlespace}
\begin{tabular}{ l 
c c c c c c c c c
 } \hline \hline
\\
 & \textbf{Kerry} & \textbf{Edwards} & \textbf{Clark} & \textbf{Dean} & \textbf{Gepardt} & \textbf{Lieberman} & \textbf{Kucinich} & \textbf{Mosely-Braun} & \textbf{Sharpton} \\ 

\\
\textbf{Baseline} & 17.40\% & 7.33\% & 15.27\% & 26.16\% & 10.16\% & 15.52\% & 2.10\% & 2.43\% & 3.62\% \\ 

\\
\textbf{Positive Ads} (+1000 GRPs) & 1.01\% & 0.48\% & 0.91\% & 1.35\% & 0.64\% & 0.92\% & 0.15\% & 0.17\% & 0.25\% \\ 

\\
\textbf{Negative Ads} (+1000 GRPs)\\
\hspace{.3cm}Attack Kerry &  & 1.51\% & 2.79\% & 4.00\% & 2.01\% & 2.83\% & 0.47\% & 0.54\% & 0.78\% \\ 

\hspace{.3cm}Attack Edwards & 3.30\% &  & 2.99\% & 4.33\% & 2.14\% & 3.03\% & 0.49\% & 0.57\% & 0.83\% \\ 

\hspace{.3cm}Attack Clark & 3.13\% & 1.53\% &  & 4.07\% & 2.03\% & 2.87\% & 0.47\% & 0.54\% & 0.79\% \\ 

\hspace{.3cm}Attack Dean & 2.89\% & 1.42\% & 2.62\% &  & 1.89\% & 2.65\% & 0.44\% & 0.51\% & 0.74\% \\ 

\hspace{.3cm}Attack Gepardt & 3.24\% & 1.58\% & 2.94\% & 4.24\% &  & 2.97\% & 0.49\% & 0.56\% & 0.82\% \\ 

\hspace{.3cm}Attack Lieberman & 3.12\% & 1.53\% & 2.83\% & 4.06\% & 2.03\% &  & 0.47\% & 0.54\% & 0.79\% \\ 

\hspace{.3cm}Attack Kucinich & 3.42\% & 1.66\% & 3.10\% & 4.50\% & 2.21\% & 3.14\% &  & 0.59\% & 0.86\% \\ 

\hspace{.3cm}Attack Mosely-Braun & 3.41\% & 1.66\% & 3.09\% & 4.49\% & 2.21\% & 3.13\% & 0.51\% &  & 0.86\% \\ 

\hspace{.3cm}Attack Sharpton & 3.39\% & 1.64\% & 3.07\% & 4.45\% & 2.19\% & 3.11\% & 0.51\% & 0.58\% &  \\ 

\\
\hline \hline
\end{tabular}
\end{singlespace}
\caption{\footnotesize{Effect Sizes for Advertising (2004 Democratic Primary) -- Analyses report the change in predicted vote share from increasing positive or negative ad spending by 1000 GRPs, relative to a baseline of no ad spending by all candidates.}}
\label{EffectTable3}
\end{table}
\end{landscape}
\begin{table}[htp]\centering\scriptsize
\begin{singlespace}
\begin{tabular}{ l 
c c c c c c
 } \hline \hline
\\
 & \textbf{Obama} & \textbf{Clinton} & \textbf{Edwards} & \textbf{Gravel} & \textbf{Kucinich} & \textbf{Richardson} \\ 

\\
\textbf{Baseline} & 38.89\% & 49.60\% & 9.56\% & 0.65\% & 0.85\% & 0.46\% \\ 

\\
\textbf{Positive Ads} (+1000 GRPs) & 1.64\% & 1.72\% & 0.61\% & 0.05\% & 0.06\% & 0.03\% \\ 

\\
\textbf{Negative Ads} (+1000 GRPs)\\
\hspace{.3cm}Attack Obama &  & 3.49\% & 1.64\% & 0.13\% & 0.17\% & 0.09\% \\ 

\hspace{.3cm}Attack Clinton & 3.27\% &  & 1.51\% & 0.12\% & 0.16\% & 0.08\% \\ 

\hspace{.3cm}Attack Edwards & 5.02\% & 5.07\% &  & 0.15\% & 0.20\% & 0.11\% \\ 

\hspace{.3cm}Attack Gravel & 5.43\% & 5.57\% & 2.12\% &  & 0.21\% & 0.11\% \\ 

\hspace{.3cm}Attack Kucinich & 5.42\% & 5.55\% & 2.12\% & 0.16\% &  & 0.11\% \\ 

\hspace{.3cm}Attack Richardson & 5.43\% & 5.58\% & 2.12\% & 0.16\% & 0.21\% &  \\ 

\\
\hline \hline
\end{tabular}
\end{singlespace}
\caption{\footnotesize{Effect Sizes for Advertising (2008 Democratic Primary) -- Analyses report the change in predicted vote share from increasing positive or negative ad spending by 1000 GRPs, relative to a baseline of no ad spending by all candidates.}}
\label{EffectTable4}
\end{table}
\begin{landscape}
\begin{table}[htp]\centering\scriptsize
\begin{singlespace}
\begin{tabular}{ l 
c c c c c c c c
 } \hline \hline
\\
 & \textbf{McCain} & \textbf{Romney} & \textbf{Giuliani} & \textbf{Huckabee} & \textbf{Huntsman} & \textbf{Thompson} & \textbf{Ron Paul} & \textbf{Keyes} \\ 

\\
\textbf{Baseline} & 41.59\% & 12.98\% & 10.01\% & 23.62\% & 0.67\% & 5.42\% & 4.08\% & 1.63\% \\ 

\\
\textbf{Positive Ads} (+1000 GRPs) & 1.68\% & 0.80\% & 0.64\% & 1.26\% & 0.05\% & 0.36\% & 0.28\% & 0.11\% \\ 

\\
\textbf{Negative Ads} (+1000 GRPs)\\
\hspace{.3cm}Attack McCain &  & 2.06\% & 1.67\% & 3.09\% & 0.13\% & 0.97\% & 0.75\% & 0.31\% \\ 

\hspace{.3cm}Attack Romney & 4.91\% &  & 2.04\% & 3.90\% & 0.15\% & 1.18\% & 0.90\% & 0.37\% \\ 

\hspace{.3cm}Attack Giuliani & 5.06\% & 2.58\% &  & 3.99\% & 0.16\% & 1.20\% & 0.92\% & 0.38\% \\ 

\hspace{.3cm}Attack Huckabee & 4.41\% & 2.35\% & 1.90\% &  & 0.14\% & 1.10\% & 0.84\% & 0.35\% \\ 

\hspace{.3cm}Attack Huntsman & 5.51\% & 2.75\% & 2.21\% & 4.27\% &  & 1.27\% & 0.97\% & 0.40\% \\ 

\hspace{.3cm}Attack Thompson & 5.28\% & 2.66\% & 2.14\% & 4.13\% & 0.16\% &  & 0.95\% & 0.39\% \\ 

\hspace{.3cm}Attack Ron Paul & 5.34\% & 2.69\% & 2.16\% & 4.17\% & 0.16\% & 1.24\% &  & 0.39\% \\ 

\hspace{.3cm}Attack Keyes & 5.46\% & 2.73\% & 2.19\% & 4.24\% & 0.16\% & 1.26\% & 0.97\% &  \\ 

\\
\hline \hline
\end{tabular}
\end{singlespace}
\caption{\footnotesize{Effect Sizes for Advertising (2008 Republican Primary) -- Analyses report the change in predicted vote share from increasing positive or negative ad spending by 1000 GRPs, relative to a baseline of no ad spending by all candidates.}}
\label{EffectTable5}
\end{table}
\end{landscape}
\begin{landscape}
\begin{table}[htp]\centering\scriptsize
\begin{singlespace}
\begin{tabular}{ l 
c c c c c c c c
 } \hline \hline
\\
 & \textbf{Romney} & \textbf{Gingrich} & \textbf{Paul} & \textbf{Perry} & \textbf{Huntsman} & \textbf{Santorum} & \textbf{Bachman} & \textbf{Cain} \\ 

\\
\textbf{Baseline} & 33.98\% & 24.81\% & 9.72\% & 7.26\% & 3.69\% & 15.73\% & 2.39\% & 2.43\% \\ 

\\
\textbf{Positive Ads} (+1000 GRPs) & 1.56\% & 1.30\% & 0.62\% & 0.48\% & 0.25\% & 0.93\% & 0.17\% & 0.17\% \\ 

\\
\textbf{Negative Ads} (+1000 GRPs)\\
\hspace{.3cm}Attack Romney &  & 3.39\% & 1.72\% & 1.34\% & 0.72\% & 2.53\% & 0.47\% & 0.48\% \\ 

\hspace{.3cm}Attack Gingrich & 4.21\% &  & 1.84\% & 1.42\% & 0.76\% & 2.71\% & 0.50\% & 0.51\% \\ 

\hspace{.3cm}Attack Paul & 4.81\% & 4.12\% &  & 1.57\% & 0.84\% & 3.01\% & 0.55\% & 0.56\% \\ 

\hspace{.3cm}Attack Perry & 4.91\% & 4.20\% & 2.06\% &  & 0.85\% & 3.06\% & 0.56\% & 0.57\% \\ 

\hspace{.3cm}Attack Huntsman & 5.06\% & 4.31\% & 2.11\% & 1.63\% &  & 3.14\% & 0.57\% & 0.58\% \\ 

\hspace{.3cm}Attack Santorum & 4.57\% & 3.94\% & 1.95\% & 1.51\% & 0.81\% &  & 0.53\% & 0.54\% \\ 

\hspace{.3cm}Attack Bachman & 5.11\% & 4.35\% & 2.13\% & 1.64\% & 0.88\% & 3.16\% &  & 0.59\% \\ 

\hspace{.3cm}Attack Cain & 5.11\% & 4.35\% & 2.13\% & 1.64\% & 0.88\% & 3.16\% & 0.58\% &  \\ 

\\
\hline \hline
\end{tabular}
\end{singlespace}
\caption{\footnotesize{Effect Sizes for Advertising (2012 Republican Primary) -- Analyses report the change in predicted vote share from increasing positive or negative ad spending by 1000 GRPs, relative to a baseline of no ad spending by all candidates.}}
\label{EffectTable6}
\end{table}
\end{landscape}
\begin{table}[htp]\centering\scriptsize
\begin{singlespace}
\begin{tabular}{ l 
c c
 } \hline \hline
\\
 & \textbf{Gore} & \textbf{Bradley} \\ 

\\
\textbf{Baseline} & 66.90\% & 33.10\% \\ 

\\
\textbf{Negative Ads} (+1000 GRPs)\\
\hspace{.3cm}Attack Gore &  & 2.62\% \\ 

\hspace{.3cm}Attack Bradley & 2.51\% &  \\ 

\\
\hline \hline
\end{tabular}
\end{singlespace}
\caption{\footnotesize{Effect Sizes for Advertising on Attacked Candidate (2000 Democratic Primary) -- Analyses report the change in predicted vote share from increasing negative ad spending by 1000 GRPs, relative to a baseline of no ad spending by all candidates.}}
\label{EffectTable7}
\end{table}
\begin{table}[htp]\centering\scriptsize
\begin{singlespace}
\begin{tabular}{ l 
c c c c c
 } \hline \hline
\\
 & \textbf{Bush} & \textbf{McCain} & \textbf{Forbes} & \textbf{Keyes} & \textbf{Bauer} \\ 

\\
\textbf{Baseline} & 63.78\% & 27.12\% & 4.26\% & 3.92\% & 0.93\% \\ 

\\
\textbf{Negative Ads} (+1000 GRPs)\\
\hspace{.3cm}Attack Bush &  & 2.62\% & 0.66\% & 0.61\% & 0.15\% \\ 

\hspace{.3cm}Attack McCain & 3.22\% &  & 0.86\% & 0.79\% & 0.20\% \\ 

\hspace{.3cm}Attack Forbes & 4.73\% & 4.52\% &  & 0.92\% & 0.23\% \\ 

\hspace{.3cm}Attack Keyes & 4.75\% & 4.53\% & 0.99\% &  & 0.23\% \\ 

\hspace{.3cm}Attack Bauer & 4.95\% & 4.63\% & 1.01\% & 0.93\% &  \\ 

\\
\hline \hline
\end{tabular}
\end{singlespace}
\caption{\footnotesize{Effect Sizes for Advertising on Attacked Candidate (2000 Republican Primary) -- Analyses report the change in predicted vote share from increasing negative ad spending by 1000 GRPs, relative to a baseline of no ad spending by all candidates.}}
\label{EffectTable8}
\end{table}
\begin{landscape}
\begin{table}[htp]\centering\scriptsize
\begin{singlespace}
\begin{tabular}{ l 
c c c c c c c c c
 } \hline \hline
\\
 & \textbf{Kerry} & \textbf{Edwards} & \textbf{Clark} & \textbf{Dean} & \textbf{Gepardt} & \textbf{Lieberman} & \textbf{Kucinich} & \textbf{Mosely-Braun} & \textbf{Sharpton} \\ 

\\
\textbf{Baseline} & 17.40\% & 7.33\% & 15.27\% & 26.16\% & 10.16\% & 15.52\% & 2.10\% & 2.43\% & 3.62\% \\ 

\\
\textbf{Negative Ads} (+1000 GRPs)\\
\hspace{.3cm}Attack Kerry &  & 1.51\% & 2.79\% & 4.00\% & 2.01\% & 2.83\% & 0.47\% & 0.54\% & 0.78\% \\ 

\hspace{.3cm}Attack Edwards & 3.30\% &  & 2.99\% & 4.33\% & 2.14\% & 3.03\% & 0.49\% & 0.57\% & 0.83\% \\ 

\hspace{.3cm}Attack Clark & 3.13\% & 1.53\% &  & 4.07\% & 2.03\% & 2.87\% & 0.47\% & 0.54\% & 0.79\% \\ 

\hspace{.3cm}Attack Dean & 2.89\% & 1.42\% & 2.62\% &  & 1.89\% & 2.65\% & 0.44\% & 0.51\% & 0.74\% \\ 

\hspace{.3cm}Attack Gepardt & 3.24\% & 1.58\% & 2.94\% & 4.24\% &  & 2.97\% & 0.49\% & 0.56\% & 0.82\% \\ 

\hspace{.3cm}Attack Lieberman & 3.12\% & 1.53\% & 2.83\% & 4.06\% & 2.03\% &  & 0.47\% & 0.54\% & 0.79\% \\ 

\hspace{.3cm}Attack Kucinich & 3.42\% & 1.66\% & 3.10\% & 4.50\% & 2.21\% & 3.14\% &  & 0.59\% & 0.86\% \\ 

\hspace{.3cm}Attack Mosely-Braun & 3.41\% & 1.66\% & 3.09\% & 4.49\% & 2.21\% & 3.13\% & 0.51\% &  & 0.86\% \\ 

\hspace{.3cm}Attack Sharpton & 3.39\% & 1.64\% & 3.07\% & 4.45\% & 2.19\% & 3.11\% & 0.51\% & 0.58\% &  \\ 

\\
\hline \hline
\end{tabular}
\end{singlespace}
\caption{\footnotesize{Effect Sizes for Advertising on Attacked Candidate (2004 Democratic Primary) -- Analyses report the change in predicted vote share from increasing negative ad spending by 1000 GRPs, relative to a baseline of no ad spending by all candidates.}}
\label{EffectTable9}
\end{table}
\end{landscape}
\begin{table}[htp]\centering\scriptsize
\begin{singlespace}
\begin{tabular}{ l 
c c c c c c
 } \hline \hline
\\
 & \textbf{Obama} & \textbf{Clinton} & \textbf{Edwards} & \textbf{Gravel} & \textbf{Kucinich} & \textbf{Richardson} \\ 

\\
\textbf{Baseline} & 38.89\% & 49.60\% & 9.56\% & 0.65\% & 0.85\% & 0.46\% \\ 

\\
\textbf{Negative Ads} (+1000 GRPs)\\
\hspace{.3cm}Attack Obama &  & 3.49\% & 1.64\% & 0.13\% & 0.17\% & 0.09\% \\ 

\hspace{.3cm}Attack Clinton & 3.27\% &  & 1.51\% & 0.12\% & 0.16\% & 0.08\% \\ 

\hspace{.3cm}Attack Edwards & 5.02\% & 5.07\% &  & 0.15\% & 0.20\% & 0.11\% \\ 

\hspace{.3cm}Attack Gravel & 5.43\% & 5.57\% & 2.12\% &  & 0.21\% & 0.11\% \\ 

\hspace{.3cm}Attack Kucinich & 5.42\% & 5.55\% & 2.12\% & 0.16\% &  & 0.11\% \\ 

\hspace{.3cm}Attack Richardson & 5.43\% & 5.58\% & 2.12\% & 0.16\% & 0.21\% &  \\ 

\\
\hline \hline
\end{tabular}
\end{singlespace}
\caption{\footnotesize{Effect Sizes for Advertising on Attacked Candidate (2008 Democratic Primary) -- Analyses report the change in predicted vote share from increasing negative ad spending by 1000 GRPs, relative to a baseline of no ad spending by all candidates.}}
\label{EffectTable10}
\end{table}
\begin{landscape}
\begin{table}[htp]\centering\scriptsize
\begin{singlespace}
\begin{tabular}{ l 
c c c c c c c c
 } \hline \hline
\\
 & \textbf{McCain} & \textbf{Romney} & \textbf{Giuliani} & \textbf{Huckabee} & \textbf{Huntsman} & \textbf{Thompson} & \textbf{Ron Paul} & \textbf{Keyes} \\ 

\\
\textbf{Baseline} & 41.59\% & 12.98\% & 10.01\% & 23.62\% & 0.67\% & 5.42\% & 4.08\% & 1.63\% \\ 

\\
\textbf{Negative Ads} (+1000 GRPs)\\
\hspace{.3cm}Attack McCain &  & 2.06\% & 1.67\% & 3.09\% & 0.13\% & 0.97\% & 0.75\% & 0.31\% \\ 

\hspace{.3cm}Attack Romney & 4.91\% &  & 2.04\% & 3.90\% & 0.15\% & 1.18\% & 0.90\% & 0.37\% \\ 

\hspace{.3cm}Attack Giuliani & 5.06\% & 2.58\% &  & 3.99\% & 0.16\% & 1.20\% & 0.92\% & 0.38\% \\ 

\hspace{.3cm}Attack Huckabee & 4.41\% & 2.35\% & 1.90\% &  & 0.14\% & 1.10\% & 0.84\% & 0.35\% \\ 

\hspace{.3cm}Attack Huntsman & 5.51\% & 2.75\% & 2.21\% & 4.27\% &  & 1.27\% & 0.97\% & 0.40\% \\ 

\hspace{.3cm}Attack Thompson & 5.28\% & 2.66\% & 2.14\% & 4.13\% & 0.16\% &  & 0.95\% & 0.39\% \\ 

\hspace{.3cm}Attack Ron Paul & 5.34\% & 2.69\% & 2.16\% & 4.17\% & 0.16\% & 1.24\% &  & 0.39\% \\ 

\hspace{.3cm}Attack Keyes & 5.46\% & 2.73\% & 2.19\% & 4.24\% & 0.16\% & 1.26\% & 0.97\% &  \\ 

\\
\hline \hline
\end{tabular}
\end{singlespace}
\caption{\footnotesize{Effect Sizes for Advertising on Attacked Candidate (2008 Republican Primary) -- Analyses report the change in predicted vote share from increasing negative ad spending by 1000 GRPs, relative to a baseline of no ad spending by all candidates.}}
\label{EffectTable11}
\end{table}
\end{landscape}
\begin{landscape}
\begin{table}[htp]\centering\scriptsize
\begin{singlespace}
\begin{tabular}{ l 
c c c c c c c c
 } \hline \hline
\\
 & \textbf{Romney} & \textbf{Gingrich} & \textbf{Paul} & \textbf{Perry} & \textbf{Huntsman} & \textbf{Santorum} & \textbf{Bachman} & \textbf{Cain} \\ 

\\
\textbf{Baseline} & 33.98\% & 24.81\% & 9.72\% & 7.26\% & 3.69\% & 15.73\% & 2.39\% & 2.43\% \\ 

\\
\textbf{Negative Ads} (+1000 GRPs)\\
\hspace{.3cm}Attack Romney &  & 3.39\% & 1.72\% & 1.34\% & 0.72\% & 2.53\% & 0.47\% & 0.48\% \\ 

\hspace{.3cm}Attack Gingrich & 4.21\% &  & 1.84\% & 1.42\% & 0.76\% & 2.71\% & 0.50\% & 0.51\% \\ 

\hspace{.3cm}Attack Paul & 4.81\% & 4.12\% &  & 1.57\% & 0.84\% & 3.01\% & 0.55\% & 0.56\% \\ 

\hspace{.3cm}Attack Perry & 4.91\% & 4.20\% & 2.06\% &  & 0.85\% & 3.06\% & 0.56\% & 0.57\% \\ 

\hspace{.3cm}Attack Huntsman & 5.06\% & 4.31\% & 2.11\% & 1.63\% &  & 3.14\% & 0.57\% & 0.58\% \\ 

\hspace{.3cm}Attack Santorum & 4.57\% & 3.94\% & 1.95\% & 1.51\% & 0.81\% &  & 0.53\% & 0.54\% \\ 

\hspace{.3cm}Attack Bachman & 5.11\% & 4.35\% & 2.13\% & 1.64\% & 0.88\% & 3.16\% &  & 0.59\% \\ 

\hspace{.3cm}Attack Cain & 5.11\% & 4.35\% & 2.13\% & 1.64\% & 0.88\% & 3.16\% & 0.58\% &  \\ 

\\
\hline \hline
\end{tabular}
\end{singlespace}
\caption{\footnotesize{Effect Sizes for Advertising on Attacked Candidate (2012 Republican Primary) -- Analyses report the change in predicted vote share from increasing negative ad spending by 1000 GRPs, relative to a baseline of no ad spending by all candidates.}}
\label{EffectTable12}
\end{table}
\end{landscape}
\begin{table}[htp]\centering\scriptsize
\begin{singlespace}
\begin{tabular}{ l 
c c c c c c
 } \hline \hline
\\ \textbf{DV:} &\multicolumn{4}{c}{\textbf{Favorability}} & \multicolumn{2}{c}{\textbf{Vote Intention}} \\
\\
 & Observations & Per. w/   & Respondents & Per. w/   & Respondents & Per. w/   \\
 &      & Pos. Adv. &       & Pos. Adv. &       & Pos. Adv. \\
\\
Full Sample & 566500 & 10.5 & 198123 & 9.7 & 121015 & 27.1 \\ 

\\
\\ 2 Week Window: \\
\hspace{.4cm}No Eminent Election & 306273 & 9.3 & 105634 & 8.8 & 54160 & 12.7 \\ 

\hspace{.4cm}Eminent Election & 26860 & 51.5 & 4916 & 47.6 & 13056 & 60.3 \\ 

\\
\\ 4 Week Window: \\
\hspace{.4cm}No Eminent Election & 306273 & 8.0 & 101647 & 7.7 & 42732 & 8.5 \\ 

\hspace{.4cm}Eminent Election & 26860 & 46.3 & 8991 & 43.0 & 24484 & 45.4 \\ 

\\
\\ 8 Week Window: \\
\hspace{.4cm}No Eminent Election & 277205 & 6.3 & 93039 & 5.9 & 28740 & 7.2 \\ 

\hspace{.4cm}Eminent Election & 55928 & 34.9 & 17626 & 34.5 & 38476 & 32.9 \\ 

\\
After Election & 233367 & 9.6 & 88521 & 9.4 & 53799 & 33.5 \\ 

\\
\\
\hline \hline
\end{tabular}
\end{singlespace}
\caption{\footnotesize{Sample Summary}}
\label{SampTable}
\end{table}
\begin{table}[htp]\centering\scriptsize
\begin{singlespace}
\begin{tabular}{ l 
c c c c c c
 } \hline \hline
\\
 & (1) & (2) & (3) & (4) & (5) & (6)
\\ \textbf{DV:} & \multicolumn{3}{c}{Favorability} & \multicolumn{3}{c}{Candidate Choice} \\
 & 2 weeks out & 4 weeks out & 8 weeks out & 2 weeks out & 4 weeks out & 8 weeks out \\ 

 \\
\textbf{Independent Variables:}\\
\hspace{.4cm}Party Match & 0.451*** & 0.460*** & 0.493*** &  &  &  \\ 

 & (0.003) & (0.003) & (0.003) &  &  &  \\ 

\hspace{.4cm}Run Positive Ads & 0.042*** & 0.047*** & 0.057*** & 0.736*** & 0.528*** & 0.733*** \\ 

 & (0.007) & (0.008) & (0.012) & (0.085) & (0.129) & (0.188) \\ 

\hspace{.4cm}Run Negative Ads & 0.091*** & 0.113*** & 0.087+ & 1.912*** & 1.790*** & 1.233*** \\ 

 & (0.025) & (0.028) & (0.048) & (0.217) & (0.241) & (0.344) \\ 

\hspace{.4cm}Attacked by Negative Ads & 0.033 & 0.084** & 0.082 & 0.795*** & 1.132*** & 1.206** \\ 

 & (0.027) & (0.032) & (0.052) & (0.225) & (0.255) & (0.390) \\ 

\\
\textbf{Observations} & 319044 & 306273 & 277205 &  &  &  \\ 

\textbf{Respondents} & 105679 & 101692 & 93084 & 54160 & 42732 & 28740 \\ 

\\
\hline \hline
\end{tabular}
\end{singlespace}
\caption{\footnotesize{The Effects of Ads on Favorability and Candidate Choice, Independent Variables Based on Number of Ads. Analyses include respondent and candidate fixed effects (columns 1-3) and candidate fixed effects (columns 4-6). Standard errors are in parentheses and are clustered by respondent in columns 1-3. One star indicates statistical significance at the 5\% level. Two stars indicates statistical significance at the 1\% level. Three stars indicates statistical significance at the 0.1\% level. A plus sign indicates statistical significance at the 10\% level.}}
\label{ResultsTableNum}
\end{table}
\begin{table}[htp]\centering\scriptsize
\begin{singlespace}
\begin{tabular}{ l 
c c c c c c
 } \hline \hline
\\
 & (1) & (2) & (3) & (4) & (5) & (6)
\\ \textbf{DV:} & \multicolumn{3}{c}{Favorability} & \multicolumn{3}{c}{Candidate Choice} \\
 & 2 weeks out & 4 weeks out & 8 weeks out & 2 weeks out & 4 weeks out & 8 weeks out \\ 

 \\
\textbf{Independent Variables:}\\
\hspace{.4cm}Party Match & 0.449*** & 0.458*** & 0.492*** &  &  &  \\ 

 & (0.003) & (0.003) & (0.003) &  &  &  \\ 

\hspace{.4cm}Run Positive Ads & 0.005** & 0.005*** & 0.006** & 0.047** & 0.020 & 0.021 \\ 

 & (0.002) & (0.002) & (0.002) & (0.016) & (0.022) & (0.033) \\ 

\hspace{.4cm}Run Negative Ads & 0.018*** & 0.018*** & 0.013+ & 0.271*** & 0.238*** & 0.130* \\ 

 & (0.004) & (0.004) & (0.007) & (0.039) & (0.047) & (0.060) \\ 

\hspace{.4cm}Attacked by Negative Ads & 0.000 & 0.002 & 0.011+ & 0.100** & 0.095** & 0.077+ \\ 

 & (0.003) & (0.004) & (0.006) & (0.031) & (0.036) & (0.045) \\ 

\\
\textbf{Observations} & 319044 & 306273 & 277205 &  &  &  \\ 

\textbf{Respondents} & 105898 & 101911 & 93296 & 54160 & 42732 & 28740 \\ 

\\
\hline \hline
\end{tabular}
\end{singlespace}
\caption{\footnotesize{The Effects of Ads on Favorability and Candidate Choice, Candidate-Month Fixed Effects. Analyses include respondent and candidate fixed effects (columns 1-3) and candidate fixed effects (columns 4-6). Standard errors are in parentheses and are clustered by respondent in columns 1-3. One star indicates statistical significance at the 5\% level. Two stars indicates statistical significance at the 1\% level. Three stars indicates statistical significance at the 0.1\% level. A plus sign indicates statistical significance at the 10\% level.}}
\label{ResultsTableCM}
\end{table}
\begin{table}[htp]\centering\scriptsize
\begin{singlespace}
\begin{tabular}{ l 
c c c c c c
 } \hline \hline
\\
 & (1) & (2) & (3) & (4) & (5) & (6)
\\ \textbf{DV:} & \multicolumn{3}{c}{Favorability} & \multicolumn{3}{c}{Candidate Choice} \\
 & 2 weeks out & 4 weeks out & 8 weeks out & 2 weeks out & 4 weeks out & 8 weeks out \\ 

 \\
\textbf{Independent Variables:}\\
\hspace{.4cm}Party Match & 0.451*** & 0.460*** & 0.493*** &  &  &  \\ 

 & (0.003) & (0.003) & (0.003) &  &  &  \\ 

\hspace{.4cm}Run Positive Ads & 0.020*** & 0.022*** & 0.018** & 0.535*** & 0.430*** & 0.588*** \\ 

 & (0.005) & (0.005) & (0.006) & (0.045) & (0.072) & (0.105) \\ 

\hspace{.4cm}Run Negative Ads & 0.042*** & 0.046*** & 0.043* & 0.514*** & 0.490*** & 0.219+ \\ 

 & (0.010) & (0.011) & (0.017) & (0.079) & (0.096) & (0.133) \\ 

\hspace{.4cm}Attacked by Negative Ads & 0.006 & 0.018 & 0.037* & 0.214** & 0.330** & 0.288* \\ 

 & (0.009) & (0.011) & (0.016) & (0.081) & (0.103) & (0.141) \\ 

\\
\textbf{Observations} & 319044 & 306273 & 277205 &  &  &  \\ 

\textbf{Respondents} & 105679 & 101692 & 93084 & 54160 & 42732 & 28740 \\ 

\\
\hline \hline
\end{tabular}
\end{singlespace}
\caption{\footnotesize{The Effects of Ads on Favorability and Candidate Choice, Logged Ads. Analyses include respondent and candidate fixed effects (columns 1-3) and candidate fixed effects (columns 4-6). Standard errors are in parentheses and are clustered by respondent in columns 1-3. One star indicates statistical significance at the 5\% level. Two stars indicates statistical significance at the 1\% level. Three stars indicates statistical significance at the 0.1\% level. A plus sign indicates statistical significance at the 10\% level.}}
\label{ResultsTableLog}
\end{table}
\begin{table}[htp]\centering\scriptsize
\begin{singlespace}
\begin{tabular}{ l 
c c c c c c
 } \hline \hline
\\
 & (1) & (2) & (3) & (4) & (5) & (6)
\\ \textbf{DV:} & \multicolumn{3}{c}{Favorability} & \multicolumn{3}{c}{Candidate Choice} \\
 & 2 weeks out & 4 weeks out & 8 weeks out & 2 weeks out & 4 weeks out & 8 weeks out \\ 

 \\
\textbf{Independent Variables:}\\
\hspace{.4cm}Party Match & 0.446*** & 0.455*** & 0.489*** &  &  &  \\ 

 & (0.003) & (0.003) & (0.003) &  &  &  \\ 

\hspace{.4cm}Run Positive Ads & 0.008*** & 0.010*** & 0.009*** & 0.117*** & 0.061* & 0.082** \\ 

 & (0.002) & (0.002) & (0.002) & (0.016) & (0.024) & (0.032) \\ 

\hspace{.4cm}Run Negative Ads & 0.015*** & 0.014*** & 0.007 & 0.283*** & 0.219*** & 0.098+ \\ 

 & (0.004) & (0.004) & (0.007) & (0.041) & (0.045) & (0.054) \\ 

\hspace{.4cm}Attacked by Negative Ads & 0.000 & 0.005 & 0.007 & 0.096** & 0.108** & 0.080+ \\ 

 & (0.004) & (0.005) & (0.007) & (0.032) & (0.034) & (0.041) \\ 

\\
\textbf{Observations} & 306207 & 294048 & 266169 &  &  &  \\ 

\textbf{Respondents} & 101490 & 97709 & 89520 & 53665 & 42382 & 28552 \\ 

\\
\hline \hline
\end{tabular}
\end{singlespace}
\caption{\footnotesize{The Effects of Ads on Favorability and Candidate Choice, No IA, NH, or SC. Analyses include respondent and candidate fixed effects (columns 1-3) and candidate fixed effects (columns 4-6). Standard errors are in parentheses and are clustered by respondent in columns 1-3. One star indicates statistical significance at the 5\% level. Two stars indicates statistical significance at the 1\% level. Three stars indicates statistical significance at the 0.1\% level. A plus sign indicates statistical significance at the 10\% level.}}
\label{ResultsTableNoEarly}
\end{table}
\begin{table}[htp]\centering\scriptsize
\begin{singlespace}
\begin{tabular}{ l 
c c c c c c
 } \hline \hline
\\
 & (1) & (2) & (3) & (4) & (5) & (6)
\\ \textbf{DV:} & \multicolumn{3}{c}{Favorability} & \multicolumn{3}{c}{Candidate Choice} \\
 & 2 weeks out & 4 weeks out & 8 weeks out & 2 weeks out & 4 weeks out & 8 weeks out \\ 

 \\
\textbf{Independent Variables:}\\
\hspace{.4cm}Party Match & 0.451*** & 0.460*** & 0.493*** &  &  &  \\ 

 & (0.003) & (0.003) & (0.003) &  &  &  \\ 

\hspace{.4cm}Run Positive Ads & 0.009*** & 0.010*** & 0.008* & 0.183*** & 0.110** & 0.098* \\ 

 & (0.003) & (0.003) & (0.003) & (0.021) & (0.036) & (0.048) \\ 

\hspace{.4cm}Run Negative Ads & 0.029*** & 0.031*** & 0.024* & 0.389*** & 0.339*** & 0.183* \\ 

 & (0.006) & (0.007) & (0.010) & (0.059) & (0.064) & (0.077) \\ 

\hspace{.4cm}Attacked by Negative Ads & 0.004 & 0.008 & 0.018+ & 0.161*** & 0.152** & 0.123+ \\ 

 & (0.005) & (0.006) & (0.009) & (0.048) & (0.052) & (0.066) \\ 

\\
\textbf{Observations} & 319044 & 306273 & 277205 &  &  &  \\ 

\textbf{Respondents} & 105679 & 101692 & 93084 & 54160 & 42732 & 28740 \\ 

\\
\hline \hline
\end{tabular}
\end{singlespace}
\caption{\footnotesize{The Effects of Ads on Favorability and Candidate Choice, Independent Variables Based on Two Weeks of Ads. Analyses include respondent and candidate fixed effects (columns 1-3) and candidate fixed effects (columns 4-6). Standard errors are in parentheses and are clustered by respondent in columns 1-3. One star indicates statistical significance at the 5\% level. Two stars indicates statistical significance at the 1\% level. Three stars indicates statistical significance at the 0.1\% level. A plus sign indicates statistical significance at the 10\% level.}}
\label{ResultsTableWk2}
\end{table}
\begin{table}[htp]\centering\scriptsize
\begin{singlespace}
\begin{tabular}{ l 
c c c c c c
 } \hline \hline
\\
 & (1) & (2) & (3) & (4) & (5) & (6)
\\ \textbf{DV:} & \multicolumn{3}{c}{Favorability} & \multicolumn{3}{c}{Candidate Choice} \\
 & 2 weeks out & 4 weeks out & 8 weeks out & 2 weeks out & 4 weeks out & 8 weeks out \\ 

 \\
\textbf{Independent Variables:}\\
\hspace{.4cm}Party Match & 0.451*** & 0.460*** & 0.493*** &  &  &  \\ 

 & (0.003) & (0.003) & (0.003) &  &  &  \\ 

\hspace{.4cm}Run Positive Ads & 0.004*** & 0.004*** & 0.003** & 0.094*** & 0.072*** & 0.079*** \\ 

 & (0.001) & (0.001) & (0.001) & (0.013) & (0.018) & (0.023) \\ 

\hspace{.4cm}Run Negative Ads & 0.009*** & 0.010*** & 0.007 & 0.168*** & 0.126*** & 0.027 \\ 

 & (0.002) & (0.003) & (0.005) & (0.030) & (0.034) & (0.041) \\ 

\hspace{.4cm}Attacked by Negative Ads & 0.002 & 0.004 & 0.001 & 0.078*** & 0.087*** & 0.072* \\ 

 & (0.003) & (0.003) & (0.004) & (0.022) & (0.024) & (0.028) \\ 

\\
\textbf{Observations} & 319044 & 306273 & 277205 &  &  &  \\ 

\textbf{Respondents} & 105679 & 101692 & 93084 & 54160 & 42732 & 28740 \\ 

\\
\hline \hline
\end{tabular}
\end{singlespace}
\caption{\footnotesize{The Effects of Ads on Favorability and Candidate Choice, Independent Variables Based on Eight Weeks of Ads. Analyses include respondent and candidate fixed effects (columns 1-3) and candidate fixed effects (columns 4-6). Standard errors are in parentheses and are clustered by respondent in columns 1-3. One star indicates statistical significance at the 5\% level. Two stars indicates statistical significance at the 1\% level. Three stars indicates statistical significance at the 0.1\% level. A plus sign indicates statistical significance at the 10\% level.}}
\label{ResultsTableWk8}
\end{table}
\begin{landscape}
\begin{table}[htp]\centering\scriptsize
\begin{singlespace}
\begin{tabular}{ l 
c c c c c c c c c
 } \hline \hline
\\ & \multicolumn{3}{c}{2 weeks out} & \multicolumn{3}{c}{4 weeks out} & \multicolumn{3}{c}{8 weeks out} \\
\textbf{Relative vote share} & 0 & 0.5 & 1 & 0 & 0.5 & 1 & 0 & 0.5 & 1 \\
\textbf{of Attacked Candidate} \\ \\
\\
Baseline & 0.462$\dagger$ & 0.562$\dagger$ & 0.715 & 0.305$\dagger$ & 0.403$\dagger$ & 0.592 & 0.990 & 1.699 & 5.986 \\ 

 & (0.095) & (0.118) & (0.184) & (0.127) & (0.172) & (0.296) & (0.684) & (1.671) & (20.382) \\ 

Number of Ads & 0.385$\dagger$ & 0.486$\dagger$ & 0.659$\dagger$ & 0.295$\dagger$ & 0.431$\dagger$ & 0.802 & 0.594 & 1.162 & 26.545 \\ 

 & (0.068) & (0.090) & (0.164) & (0.089) & (0.139) & (0.387) & (0.241) & (0.673) & (375.015) \\ 

Cand.-Month FEs & 0.174$\dagger$ & 0.213$\dagger$ & 0.274$\dagger$ & 0.082$\dagger$ & 0.103$\dagger$ & 0.137$\dagger$ & 0.164$\dagger$ & 0.233 & 0.401 \\ 

 & (0.067) & (0.081) & (0.110) & (0.097) & (0.121) & (0.163) & (0.274) & (0.401) & (0.783) \\ 

Log & 1.041 & 1.315 & 1.785 & 0.877 & 1.323 & 2.689 & 2.681 & 7.807 & -8.559 \\ 

 & (0.198) & (0.284) & (0.586) & (0.253) & (0.461) & (1.987) & (1.808) & (14.012) & (20.702) \\ 

No IA, NH, or SC & 0.414$\dagger$ & 0.499$\dagger$ & 0.626$\dagger$ & 0.277$\dagger$ & 0.368$\dagger$ & 0.548 & 0.842 & 1.426 & 4.657 \\ 

 & (0.086) & (0.105) & (0.155) & (0.128) & (0.175) & (0.302) & (0.594) & (1.383) & (14.138) \\ 

IVs Based on 2 Weeks of Ads & 0.470$\dagger$ & 0.593$\dagger$ & 0.802 & 0.326$\dagger$ & 0.420$\dagger$ & 0.590 & 0.537 & 0.808 & 1.630 \\ 

 & (0.090) & (0.120) & (0.216) & (0.125) & (0.164) & (0.265) & (0.358) & (0.632) & (2.335) \\ 

IVs Based on 8 Weeks of Ads & 0.558$\dagger$ & 0.728 & 1.047 & 0.574 & 0.879 & 1.869 & 2.934 & -8.735 & -1.755 \\ 

 & (0.135) & (0.188) & (0.354) & (0.235) & (0.419) & (1.566) & (4.792) & (34.490) & (1.358) \\ 

\\
\hline \hline
\end{tabular}
\end{singlespace}
\caption{\footnotesize{Robustness Checks for the Relative Effectiveness of Positive Advertising -- Each entry computes $\frac{     \theta_1}{\theta_2 - \tfrac{e^{u_2}}{e^{u_2} + C} \theta_3}$ for a particular model, where $\tfrac{e^{u_2}}{e^{u_2} + C}$ is the vote share of the attacked candidate relative to all candidates except the attacker. A dagger indicates that the coefficient is statistically significantly different than 1 at the 5\% level.}}
\label{NegRatRobTable}
\end{table}
\end{landscape}
\begin{table}[htp]\centering\scriptsize
\begin{singlespace}
\begin{tabular}{ l 
c c c
 } \hline \hline
\\ & 2 weeks out & 4 weeks out & 8 weeks out \\ \\
Baseline & 0.354*$\dagger$ & 0.484*$\dagger$ & 0.835 \\ 

 & (0.110) & (0.138) & (0.496) \\ 

Number of Ads & 0.416*$\dagger$ & 0.632*$\dagger$ & 0.978* \\ 

 & (0.109) & (0.133) & (0.313) \\ 

Cand.-Month FEs & 0.367*$\dagger$ & 0.398*$\dagger$ & 0.591 \\ 

 & (0.102) & (0.131) & (0.320) \\ 

Log & 0.417*$\dagger$ & 0.674* & 1.313 \\ 

 & (0.153) & (0.209) & (0.879) \\ 

No IA, NH, or SC & 0.339*$\dagger$ & 0.495*$\dagger$ & 0.819 \\ 

 & (0.102) & (0.141) & (0.475) \\ 

IVs Based on 2 Weeks of Ads & 0.413*$\dagger$ & 0.448*$\dagger$ & 0.671 \\ 

 & (0.111) & (0.137) & (0.367) \\ 

IVs Based on 8 Weeks of Ads & 0.466*$\dagger$ & 0.693* & 2.672 \\ 

 & (0.118) & (0.185) & (3.641) \\ 

\\
\hline \hline
\end{tabular}
\end{singlespace}
\caption{\footnotesize{Robustness Checks for the Crossover Point. Each entry computes $\frac{\theta_3}{\theta_2}$ for a particular model, where attacked candidates loose vote share when $\tfrac{1}{1+C} < \frac{\theta_3}{\theta_2}$ and gain vote share otherwise. Here, $\tfrac{1}{1+C}$ is the vote share of the attacking candidate relative to all candidates except the attacked candidate. One star indicates that the coefficient is statistically significantly different from 0 at the 5\% level and one dagger indicates that the coefficient is statistically significantly different from 1 at the 5\% level.}}
\label{CrossRobTable}
\end{table}
\begin{table}[htp]\centering\scriptsize
\begin{singlespace}
\begin{tabular}{ l 
c c c
 } \hline \hline
\\
 & (1) & (2) & (3)
\\ \textbf{DV:} & \multicolumn{3}{c}{Favorability} \\
 & 2 weeks out & 4 weeks out & 8 weeks out \\ 

 \\
\textbf{Independent Variables:}\\
\hspace{.4cm}Party Match & 0.509*** & 0.513*** & 0.531*** \\ 

 & (0.004) & (0.004) & (0.005) \\ 

\hspace{.4cm}Run Positive Ads & 0.006** & 0.006** & 0.005* \\ 

 & (0.002) & (0.002) & (0.002) \\ 

\hspace{.4cm}Run Negative Ads & 0.016*** & 0.017*** & 0.008 \\ 

 & (0.005) & (0.005) & (0.009) \\ 

\hspace{.4cm}Attacked by Negative Ads & 0.001 & 0.003 & 0.008 \\ 

 & (0.004) & (0.005) & (0.008) \\ 

\\
\textbf{Observations} & 301803 & 290664 & 265562 \\ 

\textbf{Respondents} & 104496 & 100628 & 92268 \\ 

\\
\hline \hline
\end{tabular}
\end{singlespace}
\caption{\footnotesize{The Effects of Ads on Favorability, Binary Measure. Analyses include respondent and candidate fixed effects. Standard errors clustered by respondents are in parentheses. One star indicates statistical significance at the 5\% level. Two stars indicates statistical significance at the 1\% level. Three stars indicates statistical significance at the 0.1\% level. A plus sign indicates statistical significance at the 10\% level.}}
\label{ResultsTableBin}
\end{table}
\begin{landscape}
\begin{table}[htp]\centering\scriptsize
\begin{singlespace}
\begin{tabular}{ l 
c c c c c c
 } \hline \hline
\\
 & (1) & (2) & (3) & (4) & (5) & (6)
\\ \textbf{DV:} & \multicolumn{3}{c}{Favorability} & \multicolumn{3}{c}{Candidate Choice} \\
 & 2 weeks out & 4 weeks out & 8 weeks out & 2 weeks out & 4 weeks out & 8 weeks out \\ 

 \\
\textbf{Independent Variables:}\\
\hspace{.4cm}Party Match & 0.451*** & 0.460*** & 0.493*** &  &  &  \\ 

 & (0.003) & (0.003) & (0.003) &  &  &  \\ 

\hspace{.4cm}Run Positive Ads & 0.005*** & 0.005** & 0.005* & 0.114*** & 0.058* & 0.090** \\ 

 & (0.001) & (0.002) & (0.002) & (0.016) & (0.024) & (0.032) \\ 

\hspace{.4cm}Run Contrast Ads & 0.016*** & 0.016*** & 0.009 & 0.325*** & 0.325*** & 0.176** \\ 

 & (0.004) & (0.004) & (0.007) & (0.038) & (0.047) & (0.060) \\ 

\hspace{.4cm}Run Negative Ads & 0.030 & 0.099* & 0.113* & 0.071 & -0.016 & -0.091 \\ 

 & (0.025) & (0.044) & (0.048) & (0.062) & (0.075) & (0.100) \\ 

\hspace{.4cm}Attacked by Contrast Ads & 0.009 & 0.010+ & 0.016 & 0.214** & 0.424*** & 0.445** \\ 

 & (0.005) & (0.006) & (0.010) & (0.067) & (0.113) & (0.156) \\ 

\hspace{.4cm}Attacked by Negative Ads & -0.006 & -0.005 & 0.001 & 0.034 & 0.029 & 0.001 \\ 

 & (0.005) & (0.007) & (0.010) & (0.036) & (0.038) & (0.052) \\ 

\\
\textbf{Observations} & 319044 & 306273 & 277205 &  &  &  \\ 

\textbf{Respondents} & 105681 & 101694 & 93086 & 54160 & 42732 & 28740 \\ 

\textbf{P-Value for Test of Attack = Contrast} & 0.567 & 0.059+ & 0.032* &  &  &  \\ 

\textbf{P-Value for Test of Get Attacked = Get Contrasted} & 0.075+ & 0.141 & 0.300 &  &  &  \\ 

\textbf{Con. - Att. Rel. Eff. at 0} &  &  &  & -1.250 & 3.803 & 1.498 \\ 

\textbf{Con. - Att. Rel. Eff. at 1/2} &  &  &  & -1.579 & 2.413 & -0.954 \\ 

\textbf{Con. - Att. Rel. Eff. at 1} &  &  &  & -2.029 & 0.701 & 0.643 \\ 

\textbf{Con. - Att. Crossover} &  &  &  & 0.183 & 3.114 & 2.537* \\ 

\\
\hline \hline
\end{tabular}
\end{singlespace}
\caption{\footnotesize{The Effects of Ads on Favorability and Candidate Choice, Contrast vs. Primarily Negative Ads. Analyses include respondent and candidate fixed effects (columns 1-3) and candidate fixed effects (columns 4-6). Standard errors are in parentheses and are clustered by respondent in columns 1-3. One star indicates statistical significance at the 5\% level. Two stars indicates statistical significance at the 1\% level. Three stars indicates statistical significance at the 0.1\% level. A plus sign indicates statistical significance at the 10\% level.}}
\label{ContrastTable}
\end{table}
\end{landscape}
\end{document}
